U.S. patent application number 15/504798 was filed with the patent office on 2017-08-24 for image processing apparatus, image processing method, recording medium, and program.
This patent application is currently assigned to RICOH COMPANY, LTD.. The applicant listed for this patent is RICOH COMPANY, LTD.. Invention is credited to Kei YASUTOMI.
Application Number | 20170244867 15/504798 |
Document ID | / |
Family ID | 55399105 |
Filed Date | 2017-08-24 |
United States Patent
Application |
20170244867 |
Kind Code |
A1 |
YASUTOMI; Kei |
August 24, 2017 |
IMAGE PROCESSING APPARATUS, IMAGE PROCESSING METHOD, RECORDING
MEDIUM, AND PROGRAM
Abstract
An image processing apparatus includes a color separation unit
separating image data into a luminance component and a color
component; a processing unit generating second image data of the
luminance component by manipulating pixel values of
multi-resolution image data generated from first image data of the
luminance component and reconstructing the manipulated
multi-resolution image data; and a composition unit compositing the
second image data of the luminance component with image data of the
color component. Further, the processing unit generates the second
image data of the luminance component based on pixel values which
are adjusted by using parameters in accordance with pixel values of
the image data of the color component.
Inventors: |
YASUTOMI; Kei; (Tokyo,
JP) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
RICOH COMPANY, LTD. |
Tokyo |
|
JP |
|
|
Assignee: |
RICOH COMPANY, LTD.
Tokyo
JP
|
Family ID: |
55399105 |
Appl. No.: |
15/504798 |
Filed: |
August 19, 2015 |
PCT Filed: |
August 19, 2015 |
PCT NO: |
PCT/JP2015/004147 |
371 Date: |
February 17, 2017 |
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
H04N 9/646 20130101;
H04N 1/6027 20130101; H04N 1/6005 20130101 |
International
Class: |
H04N 1/60 20060101
H04N001/60 |
Foreign Application Data
Date |
Code |
Application Number |
Aug 25, 2014 |
JP |
2014-170387 |
Claims
1. An image processing apparatus comprising: a color separation
unit configured to separate image data into a luminance component
and a color component; a processing unit configured to generate
second image data of the luminance component by manipulating pixel
values of multi-resolution image data generated from first image
data of the luminance component and reconstructing the manipulated
multi-resolution image data; and a composition unit configured to
composite the second image data of the luminance component with
image data of the color component, wherein the processing unit is
configured to generate the second image data of the luminance
component based on pixel values which are adjusted by using
parameters in accordance with pixel values of the image data of the
color component.
2. The image processing apparatus according to claim 1, wherein the
processing unit is configured to generate the second image data of
the luminance component by compositing the first image data of the
luminance component with third image data of the luminance
component, which is generated by the reconstructing, by using the
parameters in accordance with the pixel values of the image data of
the color component.
3. The image processing apparatus according to claim 1, wherein the
processing unit is configured to generate the second image data of
the luminance component by reconstructing a composition between the
multi-resolution image data, which is generated from the first
image data of the luminance component, and the manipulated
multi-resolution image data, in which the pixel values of the
multi-resolution image data are manipulated, by using the
parameters in accordance with the image data of the color
component.
4. The image processing apparatus according to claim 2, wherein the
parameters indicate ratios when the first image data of the
luminance component and the third image data of the luminance
component are composited with each other, and wherein the greater
the pixel values of the image data of the color component, the
greater the ratios of the first image data of the luminance
component when the first image data of the luminance component and
the third image data of the luminance component are composited with
each other.
5. The image processing apparatus according to claim 3, wherein the
parameters indicate ratios when the multi-resolution image data,
which are generated from the first image data of the luminance
component, and the manipulated multi-resolution image data, in
which the pixel values of the multi-resolution image data are
manipulated, are composited with each other and wherein the greater
the pixel values of the image data of the color component, the
greater the ratios of the multi-resolution image data generated
from the first image data of the luminance component when the
multi-resolution image data, which are generated from the first
image data of the luminance component, and the manipulated
multi-resolution image data, in which the pixel values of the
multi-resolution image data are manipulated are composited with
each other.
6. The image processing apparatus according to claim 5, wherein the
parameters are determined for each of resolution levels of
multi-resolution image data which are generated from the image data
of the color component.
7. The image processing apparatus according to claim 1, wherein the
parameters are set to one or zero when pixel values of image data
of saturation values calculated based on the image data of the
color component are less than or equal to a predetermined value
C.sub.th1 or greater than or equal to a predetermined value
C.sub.th2 (C.sub.th1<C.sub.th2).
8. The image processing apparatus according to claim 1, wherein the
processing unit is configured to generate the multi-resolution
image data of a Laplacian pyramid from the first image data of the
luminance component.
9. The image processing apparatus according to claim 1, wherein the
processing unit is configured to generate the multi-resolution
image data by performing a discrete wavelet transformation on the
first image data of the luminance component.
10. The image processing apparatus according to claim 1, wherein
the processing unit is configured to manipulate to increase pixel
values which are less than one tenth of a dynamic range of the
first image data of the luminance component and which are included
in the pixel values of the multi-resolution image data generated
from the first image data of the luminance component.
11. An image processing method comprising: a color separation step
of separating image data into a luminance component and a color
component; a processing step of generating second image data of the
luminance component by manipulating pixel values of
multi-resolution image data generated from first image data of the
luminance component and reconstructing the manipulated
multi-resolution image data; and a composition step of compositing
the second image data of the luminance component with image data of
the color component, wherein, in the processing step, the second
image data of the luminance component are generated based on pixel
values which are adjusted by using parameters in accordance with
pixel values of the image data of the color component.
12. (canceled)
13. A recording medium storing a program causing a computer to
execute an image processing method comprising: a color separation
step of separating image data into a luminance component and a
color component; a processing step of generating second image data
of the luminance component by manipulating pixel values of
multi-resolution image data generated from first image data of the
luminance component and reconstructing the manipulated
multi-resolution image data; and a composition step of compositing
the second image data of the luminance component with image data of
the color component, wherein, in the processing step, the second
image data of the luminance component are generated based on pixel
values which are adjusted by using parameters in accordance with
pixel values of the image data of the color component.
Description
TECHNICAL FIELD
[0001] The present invention relates to an image processing
apparatus, an image processing method, a recording medium, and a
program.
BACKGROUND ART
[0002] Conventionally, there have been various image processing
methods for the image data which are captured by an imaging device
such as digital camera. For example, Patent Documents 1 and 2
described below propose image processing methods in which image
data are separated into a luminance component and a color
component, then pixel values of multi-resolution image data, which
are generated based on the luminance component image data, are
manipulated and then reconstructed, and color composition is done
with the color component image data.
[0003] According to such image processing methods, it is possible
to improve problems of a lack of intelligibility and a lack of a
sense of unevenness. Here, the image data having lacked
intelligibility refers to the image data which give a hazy or foggy
impression when compared with the impression recognized when a
human actually sees. Further, the image data having lacked sense of
unevenness refers to the image data which give a planer impression
such that a sense of rampant trees or a sense of jagged rocks in an
actual scene fades away.
SUMMARY OF INVENTION
Technical Problem
[0004] However, in the image processing methods like Patent
Documents 1 and 2 where the pixel values of the multi-resolution
image data are uniformly manipulated, an unnatural luminance change
may occur in an area where a specific color component exists.
Specifically, in an area where saturation is high, an uneven
luminance which was not perceived before the image processing
(hereinafter referred to as "pseudo contour") may newly occur by
performing the image processing.
[0005] The present invention is made in light of the problem, and
an object of the present invention is to improve image quality in
the image processing which manipulates the pixel values of the
multi-resolution image data.
Solution to Problem
[0006] According to an aspect of the present invention, an image
processing apparatus includes a color separation unit separating
image data into a luminance component and a color component; a
processing unit generating second image data of the luminance
component by manipulating pixel values of multi-resolution image
data generated from first image data of the luminance component and
reconstructing the manipulated multi-resolution image data; and a
composition unit compositing the second image data of the luminance
component with image data of the color component. Further, the
processing unit generates the second image data of the luminance
component based on pixel values which are adjusted by using
parameters in accordance with pixel values of the image data of the
color component.
Advantageous Effects of Invention
[0007] According to an aspect of the present invention, it becomes
possible to improve image quality in the image processing which
manipulates the pixel values of the multi-resolution image
data.
BRIEF DESCRIPTION OF DRAWINGS
[0008] FIG. 1 is a drawing illustrating an entire configuration of
an image processing apparatus.
[0009] FIG. 2 is a drawing illustrating a hardware configuration of
the image processing apparatus.
[0010] FIG. 3 is a drawing illustrating a functional configuration
of an image processing program.
[0011] FIG. 4 is a drawing illustrating an outline of image
processing by the image processing program.
[0012] FIG. 5 is a drawing illustrating a calculation method of a
saturation value.
[0013] FIG. 6 is a drawing illustrating a relationship between the
saturation value and an adjustment ratio stored in an adjustment
ratio DB.
[0014] FIG. 7 is a drawing illustrating manipulation content of
pixel values of multi-resolution image data stored in an image
manipulation DB.
[0015] FIG. 8 is a drawing illustrating a functional configuration
of an image processing program.
[0016] FIG. 9 is a drawing illustrating an outline of image
processing by the image processing program.
[0017] FIG. 10 is a drawing illustrating a calculation method of a
saturation value.
[0018] FIG. 11 is a drawing illustrating a relationship between a
color component and the adjustment ratio stored in the adjustment
ratio DB.
[0019] FIG. 12 is a drawing illustrating the manipulation content
of the pixel values of the multi-resolution image data stored in
the image manipulation DB.
[0020] FIG. 13 is an enlarged view illustrating the manipulation
content of the pixel values of the multi-resolution image data
stored in the image manipulation DB.
DESCRIPTION OF EMBODIMENTS
[0021] In the following, embodiments of the present invention are
described with reference to the accompanying drawings. Note that
throughout the descriptions and the drawings of the embodiments
described herein, the same reference numerals are used to describe
the elements having substantially the same function as each other,
and repeated descriptions thereof may be omitted.
First Embodiment
[0022] 1. Entire Configuration of an Image Processing Apparatus
[0023] First, a whole configuration of an image processing
apparatus is described. FIG. 1 illustrates an entire configuration
of an image processing apparatus 100. As illustrated in FIG. 1, the
image processing apparatus 100 includes an image processing program
110, an adjustment ratio database (hereinafter, the term "database"
may be simplified as a "DB") 121, and an image manipulation DB
122.
[0024] The image processing program 110 performs processes of
separating the image data, which are captured by an imaging device
such as a digital camera, etc., into a luminance component and a
color component thereof and manipulating the pixel values of the
multi-resolution image data which are generated based on the
luminance component image data. Further, the image processing
program 110 reconstructs the luminance component image data based
on the multi-resolution image data whose pixel values have been
manipulated, and performs color composition with the color
component image data.
[0025] Further, in the image processing program 110, when the
luminance component image data are reconstructed, a parameter
(adjustment ratio) is used which varies in accordance with the
pixel value of the color component image data.
[0026] The adjustment ratio DB 121 stores the values of the
adjustment ratios which are used when the image processing program
110 reconstructs the luminance component image data. Specifically,
the adjustment ratio DB 121 stores the adjustment ratios, which are
calculated based on the color component image data, relative to
each pixel value of the image data of a saturation value.
[0027] The image manipulation DB 122 stores the information related
to the manipulation amount when the image processing program 110
manipulates the pixel values of the multi-resolution image data
which are generated based on the luminance component image data.
Specifically, the image manipulation DB 122 stores the pixel values
before the manipulation of the multi-resolution image data in
association with the pixel values after the manipulation of the
multi-resolution image data.
[0028] 2. Hardware Configuration of Image Processing Apparatus
[0029] Next, a hardware configuration of the image processing
apparatus 100 is described. FIG. 2 is a drawing illustrating a
hardware configuration of the image processing apparatus 100. As
illustrated in FIG. 2, the image processing apparatus 100 includes
a Central Processing Unit (CPU) 201, a Read-Only Memory (ROM) 202,
a Random Access Memory (RAM) 203, and an input/output section 204.
Further, in the image processing apparatus 100, the CPU 201, the
ROM 202, the RAM 203, and the input/output section 204 are
connected to each other via a bus 205.
[0030] The CPU 201 is a computer to execute various programs (e.g.,
the image processing program 110) stored in the ROM 202.
[0031] The ROM 202 is a non-volatile memory. The ROM 202 stores the
various programs to be executed by the CPU 201, a boot program
which is necessary to execute the various programs, various DBs
(e.g., adjustment ratio DB 121 and the image manipulation DB 122),
etc.
[0032] The RAM 203 is a main memory such as a Dynamic Random Access
Memory (DRAM), Static Random Access Memory (SRAM), etc. The RAM 203
functions as a working area which is provided when the various
programs stored in the ROM 202 are executed by the CPU 201.
[0033] The input/output section 204 transmits and receives data
with peripheral modules. The image data, which are to be processed
by executing the image processing program 110 by the CPU 201, are
input via the input/output section 204. Further, the processed
image data are output via the input/output section 204.
[0034] Further, the image processing apparatus 100 is assumed to be
used by being embedded into an apparatus having an imaging function
or an image forming function such as an imaging apparatus like a
digital camera, etc., an image forming apparatus like a scanner,
etc., a smartphone, etc. Further, the image processing apparatus
100 may function as a stand-alone apparatus having an image editing
function like a personal computer, a portable information terminal,
etc., by connecting a user interface section. In this case, the
image processing program 110 may be stored in a portable recording
medium.
[0035] 3. Functional Configuration of the Image Processing
Program
[0036] Next, the functions are described with reference to FIGS. 3
through 7, which are realized by executing the image processing
program 110 by the CPU 201. FIG. 3 is a drawing illustrating a
functional configuration of the image processing program 110
according to a first embodiment. FIG. 4 is a drawing illustrating
an outline of the image processing by the image processing program
110 according to the first embodiment. FIG. 5 is a drawing
illustrating a method of calculating the saturation value. FIG. 6
is a drawing illustrating a relationship between the saturation
value and the adjustment ratio represented in the adjustment ratio
DB 121. FIG. 7 is a drawing illustrating manipulation content.
[0037] In the following, with reference to FIGS. 4 through 7
sequentially, the functional configuration of the image processing
program 110 in FIG. 3 is described.
[0038] As illustrated in FIG. 3, the functions which are realized
by executing the image processing program 110 by the CPU 201
include the functions realized by a color separation section 310, a
luminance component processing section 320, an adjustment section
330, and a color composition section 340.
[0039] (1) Description of the Color Separation Section 310
[0040] The color separation section 310 acquires the image data
which is input via the input/output section 204. The image data 400
of FIG. 4 is an example of the image data which is acquired by the
color separation section 310. In the following description, it is
assumed that the image data 400 has an i.times.j pixel size.
[0041] The file format of the image data 400 acquired by the color
separation section 310 is a TIF format, and a color space thereof
is an RGB color space. Further, the image data 400 acquired by the
color separation section 310 has a 16-bit pixel value for each
color per pixel. Note that, however, the file format of the image
data 400 acquired by the color separation section 310 is not
limited to the TIF format. For example, the file format thereof may
be any other file format such as a JPEG format, a PNG format, etc.
Further, the color space of the image data 400 acquired by the
color separation section 310 is not limited to the RGB color space,
and may be a color space other than the RGB color space. Further,
the data amount of the pixel value per pixel in the image data 400
acquired by the color separation section 310 is not limited to 16
bits for each color.
[0042] The color separation section 310 converts the color space of
the acquired image data 400 from RGB color space to YCbCr color
space based on the following Formula (Formula 1).
Y=0.299R+0.587G+0.114B
Cr=0.500R-0.419G-0.081B
Cb=-0.169R+-0.2G+0.500B [Math.1]
[0043] Note that, however, the color space of conversion
destination by the color separation section 310 is not limited to
the YCbCr color space, and may be any color space having a
luminance component and a color component such as La*b* color
space, HSV color space, etc.
[0044] The color separation section 310 inputs the luminance (Y)
component, which is acquired by the conversion into the YCbCr color
space, into the luminance component processing section 320.
Further, the color separation section 310 inputs the color (CrCb)
component into the adjustment section 330. The image data 410 of
FIG. 4 is luminance (Y) component image data which is acquired by
the conversion to the YCbCr color space, and the image data 460 of
FIG. 4 is the color (CrCb) component image data which is acquired
by the conversion to the YCbCr color space.
[0045] (2) Description of the Adjustment Section 330
[0046] The adjustment section 330 acquires the color (CrCb)
component image data 460 which is input from the color separation
section 310, and calculates the saturation value "C" based on the
following Formula (2).
C=(Cr.sup.2+Cb.sup.2).sup.0.5 [Math.2]
[0047] Here, when the color components "Cr" and "Cb" change in the
horizontal axis and the vertical axis, respectively, the saturation
value "C" on the same concentric circle around the origin is the
same value. For example, the saturation value "C.sub.1" is equal to
the saturation value "C.sub.2" because those values are located on
the same concentric circle.
[0048] The adjustment section 330 calculates the saturation value
"C" for each pixel in the color (CrCb) component image data 460.
The image data 470 of FIG. 4 is the saturation value (C) image data
which is calculated for each pixel by the adjustment section 330.
As illustrated in FIG. 4, the saturation value (C) image data 470
includes a high saturation area and a low saturation area.
[0049] Further, the adjustment section 330 derives an adjustment
ratio ".alpha." for each pixel by referring to the adjustment ratio
DB 121 based on the calculated saturation value "C". For example,
the adjustment ratio DB 121 stores the following Formula (3), so
that the adjustment ratio DB 121 can derive the adjustment ratio
".alpha." based on Formula (3) when the saturation value "C" is
input.
.alpha. = 1.0 ( C <= C th 1 ) = 1.0 - 10.0 .times. ( C - C th 1
) ( C th 1 < C < C th 2 ) = 0.0 ( C th 2 <= C ) [ Math . 3
] ##EQU00001##
[0050] FIG. 6 illustrates a relationship between the saturation
value "C" and the adjustment ratio ".alpha." based on Formula (3).
As is apparent from FIG. 6, according to the adjustment ratio DB
121, adjustment ratio=1 is derived relative to the pixel of low
saturation and adjustment ratio=0 is derived relative to the pixel
of high saturation.
[0051] Further, the method of deriving the adjustment ratio by the
adjustment section 330 is not limited to the method based on
Formula (3). That is the adjustment ratio may be derived based on
any formula other than Formula (3). Further, the method is not
limited to a case where the adjustment ratio is expressed by a
formula(s). For example, the adjustment ratio may be derived based
on a two-dimensional lookup table on the saturation value "C"
basis.
[0052] The adjustment ratio ".alpha." derived for each pixel by the
adjustment section 330 is input to a reconstruction section 323 of
the luminance component processing section 320. Here, the
adjustment ratio ".alpha." is derived for each pixel of the
saturation value (C) image data 470 and thus, in the following, the
adjustment ratio is described as ".alpha.(x, y)".
[0053] (3) Description of the Luminance Component Processing
Section 320
[0054] Next, the luminance component processing section 320 is
described. The luminance component processing section 320 includes
a multi-resolution image generation section 321, a multi-resolution
image manipulation section 322, and the reconstruction section
323.
[0055] (3-1) Description of the Multi-Resolution Image Generation
Section 321
[0056] The multi-resolution image generation section 321 has
functions to acquire the luminance (Y) component image data 410 and
generate the multi-resolution image data 420.
[0057] Specifically, upon acquiring the luminance (Y) component
image data 410, first, the multi-resolution image generation
section 321 performs a smoothing process using a 5.times.5 Gaussian
filter. Then, the multi-resolution image generation section 321
generates the image data, whose vertical and horizontal sizes are
half of those of the luminance (Y) component image data 410
acquired by performing the smoothing process, having the pixel data
belonging to rows and columns to which even numbers are allocated
of the luminance (Y) component image data 410.
[0058] Further, in the following, the image data that have been
generated as described above are called "scaled-down image data".
Further, the header row and the header column of the luminance (Y)
component image data 410 are called a "0th row" and a "0th column",
respectively. (That is, the multi-resolution image generation
section 321 performs processing on the rows and columns to which
even numbers are allocated of the image data first.)
[0059] The image size of the scaled-down image data generated as
described above is expressed as (i/2, j/2). The multi-resolution
image generation section 321 repeatedly performs a scaled down
processing (which is a process of generating image data whose
vertical and horizontal sizes are half of those of the scaled-down
image data by performing the smoothing process first and forming an
image having rows and columns to which even numbers are allocated)
on the scaled-down image data. As a result, the vertical and
horizontal sizes of the generated image are reduced by half (1/2
size).
[0060] After that, the multi-resolution image generation section
321 doubles the vertical and horizontal sizes of the scaled-down
image data to generate the image data having the image size (i, j)
which is the same as the size of the acquired image data. In this
case, the pixel values of the scaled down image data are allocated
to the pixels of even-numbered rows and columns (rows and columns
to which respective even numbers are allocated as the row and
column numbers). Further, the pixel values of the even-numbered
rows and columns are temporarily allocated to the pixels of the
rows and columns at least one of which an odd number is allocated.
That is, four pixels having the same pixel value are generated.
After that, the smoothing process using a 5.times.5 Gaussian filter
is performed to generate image data (hereinafter, the image data
generated as described above is called "scaled-up image data").
[0061] The multi-resolution image generation section 321 generates
image data in which a difference is calculated at each pixel
between the image data before the scaled down processing is
performed thereon and the image data which is acquired by
performing the scaled down processing and then the scaled up
processing. Hereinafter, the image data generated as described
above is called "Laplacian component image data".
[0062] The multi-resolution image generation section 321 generates
the Laplacian component image data having a first resolution level
(resolution level 0). The image data 420_0 in FIG. 4 is the
Laplacian component image data having the resolution level 0 and is
the image data having the image size=(i, j).
[0063] Next, the multi-resolution image generation section 321
acquires a difference between the further scaled-down image data
and the scaled-up image data which are generated by performing the
scaled down processing and then the scaled up processing on the
further scaled-down image data. By doing this, Laplacian component
image data 420-1 in the next resolution level are calculated.
[0064] The Laplacian component image data 420-1 is the Laplacian
component image data having a resolution level 1 and is the image
data having the image size=(i/2, j/2).
[0065] After that, the multi-resolution image generation section
321 repeats the scaled down processing, the scaled up processing,
and the process of calculating the Laplacian component image data
until the Laplacian component image data 420_d having the image
size=(2, 2) are calculated. By doing this, it becomes possible to
generate a set of Laplacian component image data from the Laplacian
component image data 420_0 having the image size=(i, j) and having
the resolution level 0 to the Laplacian component image data 420_d
having the image size=(2, 2) and having the resolution level d. The
set of the Laplacian component image data corresponds to a set of
image data of the multi-resolution image data 420 having respective
resolution levels which is a so-called "Laplacian pyramid".
[0066] In the above description, the 5.times.5 Gaussian filter is
used as a smoothing filter when the multi-resolution image
generation section 321 performs the scaled down processing and the
scaled up processing. Note that, however, any other smoothing
filter may be alternatively used. Further, without using any
smoothing filter, a so-called "interpolation process" may be used
when the scaled down processing and the scaled up processing are
performed. As specific examples of the interpolation process, there
are a bi-linear method, a bi-cubic method, etc.
[0067] Further, when the scaled down processing is repeatedly
performed, there may be a case where the image size corresponds to
an odd number depending on the luminance (Y) component image data
410. In such a case, according to this embodiment, it is assumed
that the multi-resolution image generation section 321 performs the
scaled down processing by using the rows and columns to which even
numbers are allocated. However, when the image size corresponds to
an odd number, the method of the scaled down processing is not
limited thereto.
[0068] For example, when the image size (e.g., the number of rows
is P) corresponds to an odd number, the scaled down processing may
be performed by assuming that the image size (number of rows) after
the scaled down process is (P+1)/2. In this case, the pixel values
at the pixels where both row and column have respective even
numbers before the scaled down processing is performed are
reflected on the pixels after the scaled down process is
performed.
[0069] Otherwise, when the image size (the number of rows) before
the scaled down processing corresponds to an odd number, one row
may be added first and then the scaled down processing is performed
to generate the image data having a half (1/2) size. Further, at
the stage when the luminance (Y) component image data 410 is
acquired, the image size may be expanded by 2-factorial first and
the multi-resolution image data 420 is generated.
[0070] (3-2) Description of the Multi-Resolution Image Manipulation
Section 322
[0071] The multi-resolution image manipulation section 322 is
described. The multi-resolution image manipulation section 322
manipulates the pixel values of the Laplacian component image data
420_n having the resolution level n included in the
multi-resolution image data 420 generated by the multi-resolution
image generation section 321 based on the following Formula (4)
[Math. 4]
[0072] L n ' ( x , y ) = sign * ( | L n ( x , y ) | / range )
.alpha. * range ( When | L n ( x , y ) | < range ) = sign * (
grad * ( | L n ( x , y ) | - range ) + range ) ( When range < |
L n ( x , y ) | ) ##EQU00002##
[0073] Here, the elements in Formula (4) denote as follows.
[0074] The element "L'n(x, y)" denotes the corresponding pixel
values of the Laplacian component image data after manipulation at
the resolution level n, wherer "n" is in a range from 0 to d (upper
limit determined depending on input image size). The element
"|Ln(x, y)|" denotes an absolute value (magnitude) of the pixel
values in the Laplacian component image data before manipulation at
the resolution level n. The element "sign" has a value "1" or "-1"
when the value of Ln(x, y) is positive or negative,
respectively.
[0075] Here, the element "range" is set to "0.1". This is because
the luminance (Y) component image data in YCbCr color space
corresponds to the original image and the luminance (Y) component
image data is in a range from 0.0 to 1.0. Further, the value of the
"range" depends on the dynamic range of the original image.
[0076] Here, the element ".alpha." is set to 0.6.
[0077] Further, the element "grad" is set to 0.6. (Further, the
value of "grad" depends on the dynamic range of the original
image.)
[0078] FIG. 7 is a graph of the manipulation content of FIG. 4
illustrating a relationship between the Laplacian component image
data before manipulation and the Laplacian component image data
after the manipulation calculated by using Formula 4. In FIG. 7,
the bold solid line expresses the relationship of Formula 4. As a
comparison, the dotted line in the graph expresses that the
relationship of the inclination is 1 (i.e., the pixel values of the
Laplacian component image data do not change between before and
after the manipulation).
[0079] As is apparent from FIG. 7, with respect to the pixel value
having a small absolute value (pixel value having a small absolute
value of the Laplacian component image data before manipulation),
the manipulation is performed in a manner such that the pixel value
is increased and the increase amount thereof is relatively large.
On the other hand, with respect to the pixel value having a large
absolute value (pixel value having a large absolute value of the
Laplacian component image data before manipulation), the
manipulation is performed in a manner such that the pixel value is
increased but the increase amount thereof is relatively small or
the pixel value is decreased.
[0080] Further, note that the content of the manipulation by the
multi-resolution image manipulation section 322 is not limited to
the content of the manipulation illustrated in FIG. 4, and any
other content of the manipulation may be performed. Further, the
manipulation performed by the multi-resolution image manipulation
section 322 is not limited to that performed based on a formula
such as Formula (4), and the manipulation may be performed based
on, for example, a lookup table.
[0081] Further, in Formula (4), it is set that "range=0.1" because
it is assumed that the multi-resolution image data 420 of the
luminance (Y) component image data 410 in YCbCr space is
manipulated. In the case of the multi-resolution image data 420 of
the luminance (Y) component image data 410 in YCbCr space, the
dynamic range (Dr) is in a range from 0.0 to 1.0. Therefore, it is
set that range=0.1.
[0082] In other words, in a case of the multi-resolution image data
which are based on another color space having different dynamic
range, the value of the "range" changes accordingly. For example,
in the case of the multi-resolution image data which are based on
La*b* color space, the dynamic range is from 0.0 to 100.0.
Therefore, it is to be set that range=10.
[0083] That is, the content of the manipulation by the
multi-resolution image manipulation section 322 varies depending on
whether the value of the Laplacian component image data 420_n
before the manipulation at the resolution level n satisfies the
following Formula (5).
L.sub.n/Dr<0.1 [Math.5]
[0084] Based on this, the pixel value Ln where the dynamic range
(Dr) is less than 1/10 increases, and the pixel value Ln where the
dynamic range (Dr) is greater than or equal to 1/10 decreases.
[0085] (3-3) Description of the Reconstruction Section 323
[0086] The reconstruction section 323 reconstructs luminance (Y)
component image data based on multi-resolution image data 430,
which is after manipulation, by repeatedly performing manipulation
which is opposite to the manipulation of the multi-resolution image
generation section 321. Specifically, the calculation starts with
the resolution level on the lower resolution side, and the process
of adding the pixel values generated after the scaled up processing
is repeated in each of the resolution levels, so that luminance
(Y') component image data is reconstructed. In FIG. 4, the image
data 440 is the reconstructed luminance (Y') component image
data.
[0087] Further, the calculation of the scaled up processing in the
reconstruction section 323 is based on the method same as that in
the calculation of the scaled up processing in the multi-resolution
image generation section 321. That is, the image data are generated
whose image size is the same as that before the scaled down
processing (basically, the image size having doubled in both
vertical and horizontal directions), and the pixel values of the
multi-resolution image data of a low resolution level are set to
the pixels at the row and column to which respective even numbers
are allocated. In this case, the pixel values of the even-numbered
rows and columns are temporarily allocated to the pixels of the
rows and columns at least one of which an odd number is allocated
(that is, four pixels having the same pixel value are generated).
After that, the smoothing process using a 5.times.5 Gaussian filter
is performed to generate the scaled-up image data.
[0088] Further, the reconstruction section 323 composites between
the luminance (Y') component image data 440 acquired by the
reconstruction and the luminance (Y) component image data 410
before manipulation by using the adjustment ratio .alpha.(x, y)
which is input from the adjustment section 330.
[0089] Specifically, the pixel values (Y'' (x, y)) of luminance
(Y'') component image data 450 are calculated based on the
following Formula (6) by using the adjustment ratio .alpha.(x, y)
input from the adjustment section 330.
Y''(x,y)=.alpha.(x,y)*Y'(x,y)+(1-.alpha.(x,y))*Y(x,y) [Math.6]
[0090] Due to the reflection of the adjustment ratio .alpha.(x, y)
in Formula (6), the luminance (Y'') component image data 450
becomes so that:
[0091] i) pixels values in an area corresponding to the low
saturation area of the saturation value (C) image data 470 become
the pixel values on which the manipulation relative to the
multi-resolution image data is reflected. That is, the pixel values
become similar to the pixel values of the luminance (Y') component
image data 440; and
[0092] ii) pixels values in an area corresponding to the high
saturation area of the saturation value (C) image data 470 become
the pixel values on which the manipulation relative to the
multi-resolution image data is not reflected. That is, the pixel
values become similar to the pixel values of the luminance (Y)
component image data 410.
[0093] As a result, it becomes possible to prevent the occurrence
of pseudo contour in the high saturation area caused by performing
the manipulation on the multi-resolution image data.
[0094] The reconstruction section 323 outputs the luminance (Y'')
component image data 450 calculated as described above to the color
composition section 340.
[0095] (4) Description of the Color Composition Section 340
[0096] The color composition section 340 performs color composition
between the luminance (Y'') component image data 450 output from
the reconstruction section 323 and the color (CrCb) component image
data 460 output from the color separation section 310. Further, the
color composition section 340 performs conversion from the YCrCb
color space to the RGB color space on the image data acquired by
the color composition, and outputs the image data 480 having RGB
color components.
[0097] 4. Outline
[0098] As is apparent from the above descriptions, in the image
processing of manipulating the pixel values of the multi-resolution
image data in the image processing apparatus according to this
embodiment,
[0099] i) different adjustment ratios are derived based on the
color component image data between the pixels having higher
saturation and the pixels having lower saturation; and
[0100] ii) composition is performed between the luminance component
image data before manipulation and the luminance component image
data, which is acquired by restructuring the multi-resolution image
data after manipulation, by using the derived adjustment
ratios.
[0101] By doing this, on the luminance component image data to be
composited with the color component image data,
[0102] i) in the high saturation area, the luminance component
image data before manipulation is reflected; and
[0103] ii) in the low saturation area, the luminance component
image data after manipulation is reflected.
[0104] As a result, it becomes possible to prevent the occurrence
of pseudo contour in the high saturation area caused by performing
the manipulation on the multi-resolution image data. That is, it
becomes possible to improve the image quality in the image
processing of manipulating the pixel values of the multi-resolution
image data.
Second Embodiment
[0105] In the first embodiment, the composition is made between the
luminance (Y) component image data, which is acquired by performing
the color separation on the input image data, and the luminance
(Y'') component image data, which is acquired by restructuring the
multi-resolution image data after manipulation, by using the
derived adjustment ratios.
[0106] On the other hand, in a second embodiment, the composition
is made between the multi-resolution image data before manipulation
and the multi-resolution image data after manipulation by using the
derived adjustment ratios. By doing this, in the high saturation
area, it becomes possible to generate the multi-resolution image
data on which the multi-resolution image data after manipulation
have not been reflected. Further, in the low saturation area, it
becomes possible to generate the multi-resolution image data on
which the multi-resolution image data after manipulation have been
reflected. In the following, details of the second embodiment are
described. Note that the points different from those in the first
embodiment are mainly described.
[0107] 1. Functional Configuration of an Image Processing
Apparatus
[0108] With reference to FIGS. 8 and 9, the functions are described
realized by executing an image processing program 110 according to
the second embodiment by the CPU 201. FIG. 8 is a drawing
illustrating a functional configuration of the image processing
program 110 according to the second embodiment. FIG. 9 is a drawing
illustrating an outline of the image processing by the image
processing program 110 according to the second embodiment. In the
following, the functional configuration of the image processing
program 110 of FIG. 8 is described with reference to FIG. 9
occasionally. Here, the functional configuration of FIG. 8 differs
from that described with reference to FIG. 3 in the first
embodiment in an adjustment section 830, a multi-resolution image
manipulation section 822, and a reconstruction section 823.
Therefore, in the following, the adjustment section 830, the
multi-resolution image manipulation section 822, and the
reconstruction section 823 are described in detail.
[0109] (1) Description of the Adjustment Section 830
[0110] The adjustment section 830 acquires the color (Cr) component
image data and the color (Cb) component image data from the color
separation section 310. The image data 920 of FIG. 9 is the color
(Cr) component image data and the image data 930 of FIG. 9 is the
color (Cb) component image data.
[0111] The adjustment section 830 generates respective
multi-resolution image data relative to the color (Cr) component
image data and the color (Cb) component image data. Specifically,
the adjustment section 830 generates the multi-resolution image
data by repeatedly performing only the scaled down process on the
color (Cr) component image data and the color (Cb) component image
data.
[0112] The image data 940 of FIG. 9 are the multi-resolution image
data generated based on color (Cr) component image data 920. The
image data 950 of FIG. 9 are the multi-resolution image data
generated based on color (Cb) component image data 930.
[0113] Further, the processing to generate the multi-resolution
image data 940 relative to the color (Cr) component image data 920
is similar to the process to generate the multi-resolution image
data 950 relative to the color (Cb) component image data 930. Due
to this, in the following, the processing to generate the
multi-resolution image data 940 relative to the color (Cr)
component image data 920 is described.
[0114] The adjustment section 830 performs a smoothing process on
the color (Cr) component image data 920 (image data having the
image size=(i, j)) using a 5.times.5 Gaussian filter. Next, the
adjustment section 830 generates the image data, whose vertical and
horizontal sizes are half of those of the color (Cr) component
image data 920, having (only) the rows and columns to which
corresponding even numbers are allocated in the image data on which
the smooth process has been performed. Further, similar to the
first embodiment, the header row and the header column are called
the "0th row" and the "0th column", respectively. That is, similar
to the first embodiment, the scaled down processing is performed on
the rows and columns, to which even numbers are allocated, of the
color (Cr) component image data 920.
[0115] By the above processing, the Gaussian component image data
940_1 having the first resolution level of a Gaussian pyramid
(resolution level 1) is generated. Further, the Gaussian component
image data 940_0 having a resolution level 0 is the same as the
color (Cr) component image data 920 output from the color
separation section 310.
[0116] After that, the scaled down processing is repeatedly
performed, the Gaussian component image data having respective
resolution levels are generated until the image size is (2, 2). The
pixel values of the Gaussian component image data 940_n having a
resolution level n of the color component (Cr) generated as
described above are expressed as Crn(x, y).
[0117] Similarly, the adjustment section 830 generates the Gaussian
pyramid of the multi-resolution image data 950 relative to the
color (Cb) component image data 930. Here, the generated pixel
values of the Gaussian component image data 950_n having a
resolution level n of the color component (Cb) are expressed as
Cbn(x, y).
[0118] The adjustment section 830 generates saturation value (C)
multi-resolution image data by using the multi-resolution image
data 940 and the multi-resolution image data 950 generated as
described above. Specifically, the adjustment section 830 generates
saturation value (C) multi-resolution image data by using the
Gaussian component image data 940_n having the resolution level n
relative to the color (Cr) component image data 920 and the
Gaussian component image data 950_n having the resolution level n
relative to the color (Cb) component image data 930.
[0119] In FIG. 9, the image data 960 are the saturation value (C)
multi-resolution image data calculated by the adjustment section
830 for each resolution level. Further, the pixel values Cn(x, y)
of the saturation value (C) multi-resolution image data 960 are
calculated based on the following Formula (7).
C.sub.n(x,y)=(Cr.sub.n(x,y).sup.2+Cb.sub.n(x,y).sup.2).sup.0.5
[Math.7]
[0120] Based on the pixel values Cn(x, y) of the saturation value
(C) multi-resolution image data 960 generated as described above,
the adjustment section 830 calculates the adjustment ratios based
on the following Formula (8).
[Math. 8]
[0121] .alpha. n ( x , y ) = 1.0 ( C n <= C th 1 ) = 1.0 - 10.0
.times. ( C n ( x , y ) - C th 1 ) ( C th 1 < C n < C th 2 )
= 0.0 ( C th 2 <= C n ) ##EQU00003##
[0122] In Formula (8), the ".alpha.n(x, y)" denotes the adjustment
ratio to be used when the composition is made between the pixel
values of the luminance (Y) component multi-resolution image data
having the solution level n before and after manipulation.
[0123] The adjustment section 830 inputs the adjustment ratios
.alpha.n(x, y) corresponding to the pixel values Cn(x, y) of the
saturation value (C) multi-resolution image data 960 having the
resolution level n in the multi-resolution image manipulation
section 822. Here, the relationship between the pixel values Cn(x,
y) of the saturation value (C) multi-resolution image data 960
having the resolution level n and the adjustment ratios .alpha.n(x,
y) is already described in the above first embodiment with
reference to FIG. 6, therefore, the repeated description thereof is
herein omitted.
[0124] (2) Description of the Multi-Resolution Image Manipulation
Section 822
[0125] From the multi-resolution image generation section 321, the
multi-resolution image manipulation section 822 acquires the pixel
values Ln(x, y) of the Laplacian component image data 420_n having
the resolution level n included in the multi-resolution image data
420 relative to the luminance (Y) component image data 410. Then,
the multi-resolution image manipulation section 822 manipulates the
acquired pixel values Ln(x, y) of the Laplacian component image
data 420_n based on Formula (4). By doing this, the pixel values
L'n(x, y) of the Laplacian component image data 430_n after
manipulation are obtained.
[0126] The multi-resolution image manipulation section 822 makes
composition between the pixel values L'n(x, y) of the Laplacian
component image data 430_n after manipulation and the pixel values
Ln(x, y) of the Laplacian component image data 420_n before
manipulation. During the composition, the adjustment ratios
.alpha.n(x, y) output from the adjustment section 830 are used.
[0127] Specifically, based on the following Formula (9), the
multi-resolution image manipulation section 822 calculates the
pixel values L''n(x, y) of the Laplacian component image data 910_n
having the resolution level n after composition.
L.sub.n''(x,y)=.alpha..sub.n(x,y)*L.sub.n'(x,y)+(1.0-.alpha..sub.n(x,y))-
*L.sub.n(x,y) [Math.9]
[0128] By Formula (9), the adjustment ratios .alpha.n(x, y) are
reflected, so that in luminance (Y) component multi-resolution
image data 910 after composition,
[0129] i) the pixel values in the area corresponding to the low
saturation area of the saturation value (C) multi-resolution image
data 960 are the pixel values on which the manipulation relative to
the luminance (Y) component multi-resolution image data is
reflected, that is, the pixel values in the area corresponding to
the low saturation area are the pixel values of the
multi-resolution image data after manipulation; and
[0130] ii) the pixel values in the area corresponding to the high
saturation area of the saturation value (C) multi-resolution image
data 960 are the pixel values on which the manipulation relative to
the luminance (Y) component multi-resolution image data is not
reflected, that is, the pixel values in the area corresponding to
the high saturation area are the pixel values of the
multi-resolution image data before manipulation.
[0131] As a result, it becomes possible to prevent the occurrence
of pseudo contour in the high saturation area caused by performing
the manipulation on the multi-resolution image data.
[0132] (3) Description of the Reconstruction Section 823
[0133] The reconstruction section 823 reconstructs the luminance
(Y'') component image data 450 by repeatedly performing the
manipulation, which is opposite to the manipulation of the
multi-resolution image generation section 321, on the luminance (Y)
component multi-resolution image data 910 which is output from the
multi-resolution image manipulation section 822.
[0134] The luminance (Y'') component image data 450 acquired by
restructuring the luminance (Y) component multi-resolution image
data 910 in the reconstruction section 823 is input in the color
composition section 340, so that the luminance (Y'') component
image data 450 is composited with the color (Cr) component image
data 920 and the color (Cb) component image data 930.
[0135] 2. Outline
[0136] As is apparent from the above description, in the image
processing apparatus according to this embodiment, in the image
processing of manipulating the pixel values of the multi-resolution
image data,
[0137] i) multi-resolution image data relative to color component
image data are generated and the saturation values at each of the
resolution levels are calculated, so that different adjustment
ratios are derived between the pixels having high saturation and
the pixels having low saturation; and
[0138] ii) in restructuring the multi-resolution image data
generated from the luminance component image data, the composition
is made using the multi-resolution image data after manipulation
and the multi-resolution image data before manipulation by using
the adjustment ratios.
[0139] By doing this, in the reconstructed luminance component
multi-resolution image data,
[0140] i) on the high saturation area, the luminance component
multi-resolution image data before manipulation is reflected;
and
[0141] ii) on the low saturation area, the luminance component
multi-resolution image data after manipulation is reflected.
[0142] As a result, it becomes possible to prevent the occurrence
of pseudo contour in the high saturation area caused by performing
the manipulation on the multi-resolution image data. That is, it
becomes possible to improve the image quality in the image
processing of manipulating the pixel values of the multi-resolution
image data.
Third Embodiment
[0143] In the above first and second embodiments, in calculating
the saturation value (C) image data by using the color (Cr)
component image data and the color (Cb) component image data, a
square-root of sum of squares of the pixel values of the color (Cr)
component image data and the color (Cb) component image data are
used. On the other hand, in a third embodiment, the pixel values of
the color (Cr) component image data or the pixel values of the
color (Cb) component image data, whichever is greater, are set as
the saturation values (C). That is, in the third embodiment, the
adjustment sections 330 and 830 calculate the saturation values (C)
based on the following Formula (10).
C=max(|C.sub.r|,|C.sub.b|) [Math.10]
[0144] Further, as illustrated in FIG. 10, the saturation values
(C) calculated based on the following Formula (10) are arranged on
the sides of similar quadrates having a common corner at the origin
of the coordinates where the color (Cr) component is plotted along
the horizontal axis and the color (Cb) is plotted along the
vertical axis. For example, the saturation value C.sub.1 is equal
to the saturation value C.sub.2.
[0145] In the adjustment sections 330 and 830, by calculating the
saturation values (C) based on Formula (10), it becomes possible to
reduce the processing load in calculating the saturation values
(C).
Fourth Embodiment
[0146] In the above first through third embodiments, the adjustment
ratio ".alpha." is set to one when the saturation value (C) is less
than or equal to the threshold value C.sub.th1 and adjustment ratio
".alpha." is set to zero when the saturation value (C) is greater
than or equal to the threshold value C.sub.th2, so that the
manipulation relative to the multi-resolution image data is not
reflected on the high saturation area.
[0147] On the other hand, in a fourth embodiment, when the color
(Cr, Cb) components are within a predetermined range, the
adjustment ratio ".alpha." is set to zero, and otherwise, the
adjustment ratio ".alpha." is set to one. By doing this, it becomes
possible not to reflect the manipulation relative to the
multi-resolution image data on the area having predetermined color
(Cr, Cb) components. Specifically, the following Formula (11) is
stored in the adjustment ratio DB 121, so that when the color (Cr,
Cb) components are input, the adjustment ratio ".alpha." is derived
based on the following Formula (11).
.alpha. = 0.0 0.02 <= Cr <= 0.18 and - 0.2 <= Cb <= 0.0
= 1.0 When other than above [ Math . 11 ] ##EQU00004##
[0148] Further, FIG. 11 illustrates a relationship between the
color (Cr, Cb) components of Formula (11) and the adjustment ratio
".alpha.". As apparent from FIG. 11, according to the adjustment
ratio DB 121, adjustment ratio=0 is derived relative to the pixels
having the color (Cr, Cb) components within a predetermined range,
and adjustment ratio=1 is derived relative to the pixels having the
color (Cr, Cb) components outside the predetermined range.
[0149] For example, the range of the color (Cr) component and the
color (Cb) component is a range corresponding to colors of human
skin. In this case, when the adjustment ratio ".alpha." is used
derived based on Formula (11), the manipulation relative to the
luminance (Y) component multi-resolution image data is not
reflected on the area including the colors of human skin.
[0150] Generally, when the manipulation relative to luminance (Y)
component multi-resolution image data is performed, a small density
change is increased. As a result, effects of improved sense of
intelligibility and sense of unevenness can be acquired. On the
other hand, creases and pores of human skin may be emphasized. In
the fourth embodiment, by using the adjustment ratio ".alpha."
derived based on Formula (11), it becomes possible to prevent the
increase of a small density change with respect to human skin. That
is, in the image processing of manipulating the pixel values of the
multi-resolution image data, it becomes possible to both improve a
sense of intelligibility and a sense of unevenness and prevent the
degradation of human skin.
Fifth Embodiment
[0151] In the above first through fourth embodiments, a case is
described where the multi-resolution image data generated by the
multi-resolution image generation section 321 is a Laplacian
pyramid. On the other hand, in a fifth embodiment, a case is
described where the multi-resolution image generation section 321
generates the multi-resolution image data by repeatedly performing
two-dimensional discrete wavelet transformation.
[0152] The two-dimensional discrete wavelet transformation is a
known technique. Thus, a detailed description thereof is herein
omitted. When the two-dimensional discrete wavelet transformation
is performed, by a one-time transformation process, the luminance
(Y) component image data are decomposed into one low-frequency
coefficient "LL" and three high-frequency coefficients "LH", "HL",
and "HH". The multi-resolution image generation section 321 further
performs the two-dimensional discrete wavelet transformation
repeatedly on the low-frequency coefficient "LL", one low-frequency
coefficient "LL" and three high-frequency coefficients "LH", "HL",
and "HH" in the next resolution level are calculated.
[0153] The multi-resolution image manipulation section 322 applies
the manipulation of Formula (4) on the coefficients "LH", "HL", and
"HH" in each of the resolution levels calculated as described.
Then, the reconstruction section 323 reconstructs the
multi-resolution image data and generates the luminance (Y'')
component image data 450.
[0154] As described above, by using the two-dimensional discrete
wavelet transformation in generating the multi-resolution image
data, it becomes possible to enjoy a benefit of reducing a memory
area to store the multi-resolution image data.
Sixth Embodiment
[0155] In the above first through fifth embodiments, the luminance
(Y) component multi-resolution image data are manipulated based on
the manipulation content of FIG. 7. Specifically,
[0156] i) the pixel values having small absolute values (pixel
values having small absolute values of the Laplacian component
image data) are manipulated so that the pixel values are increased
and the increase amounts are relatively large; and
[0157] ii) the pixel values having large absolute values (pixel
values having large absolute values of the Laplacian component
image data) are manipulated so that the pixel values are increased
and the increase amounts are relatively small, or the pixel values
are decreased.
[0158] On the other hand, in a sixth embodiment, the manipulation
is performed in a manner such that a set of pixel values, which are
relatively small among the pixel values whose absolute values are
small, are not changed; and the pixel values, which are not
included in the set of pixel values and which are included in the
pixel values whose absolute values are small, are increased.
[0159] FIG. 12 is a drawing illustrating the content of the
manipulation relative to the multi-resolution image data by the
multi-resolution image manipulation section 322 and 822 according
to the sixth embodiment. Further, FIG. 13 is an enlarged view of an
area in the vicinity of the origin of FIG. 12.
[0160] As illustrated in FIG. 13, in the multi-resolution image
manipulation section 322 and 822 according to the sixth embodiment,
the manipulation amounts are reduced when the absolute value of the
pixel values of the Laplacian component image data is less than or
equal to 0.04. This is because when there is a noise component in
the luminance (Y) component image data, such a noise component is
more likely to be expressed as a small slope. When a small slope is
considered in view of the Laplacian component, the small slope
corresponds to a small Laplacian component. Therefore, it becomes
possible to prevent the noise component from being emphasized by
reducing the manipulation amount when the absolute value of the
Laplacian component image data is less than or equal to 0.04.
[0161] On the other hand, in the luminance (Y) component image
data, the pixels to be emphasized are the pixels whose values are
slightly greater than those of a noise component in the Laplacian
component image data included in the luminance (Y) component
multi-resolution image data. Thus, in this embodiment, it is
intended to improve a sense of intelligibility and a sense of
unevenness by increasing the manipulation amounts relative to the
pixels whose absolute values in the Laplacian component image are
slightly greater than 0.04.
Seventh Embodiment
[0162] In the sixth embodiment, it is assumed that the
multi-resolution image manipulation section 822 applies the content
of the manipulation of FIG. 12 or 13 to all the resolution levels.
On the other hand, in a seventh embodiment, the multi-resolution
image manipulation section 822 applies the content of the
manipulation of FIG. 12 or 13 to the multi-resolution image data
having high resolution levels and multi-resolution image data
having low resolution levels.
[0163] By doing this, the manipulation content, in which the
manipulation amounts are reduced relative to the pixels whose
absolute values of the Laplacian component image data are small, is
applied to the multi-resolution image data having high resolution
levels and multi-resolution image data having low resolution
levels. On the other hand, the manipulation content, in which the
manipulation amounts are not reduced relative to the pixels whose
absolute values of the Laplacian component image data are small, is
applied to the multi-resolution image data having middle resolution
levels.
[0164] Here, the "high", the "middle", and the "low" resolutions
levels are described. For example, when it is assumed that the size
of the luminance (Y) component image data to be input is
4928.times.3280 pixels, the resolution levels generated by the
multi-resolution image generation section 321 and 822 are
resolution levels 0 through 12.
[0165] In this case, for example, the resolution levels 0 through 2
are classified as the high resolution levels, so that the
multi-resolution image manipulation sections 322 and 822 apply the
manipulation content of FIG. 12 or 13.
[0166] Further, the resolution levels 3 through 7 are classified as
the middle resolution levels, so that the multi-resolution image
manipulation sections 322 and 822 apply the manipulation content of
FIG. 7. Further, the resolution levels 8 through 12 are classified
as the low resolution levels, so that the multi-resolution image
manipulation sections 322 and 822 apply the manipulation content of
FIG. 12 or 13.
[0167] Further, the reasons why the manipulation content varies
depending on the resolution levels are described below. That is, a
noise component included in the luminance (Y) component image data
is more likely to be included in the Laplacian component image data
having high resolution levels when the multi-resolution image data
are generated. On the other hand, in the Laplacian component image
data having middle resolution levels, actually existing contrast is
included even in a very small slope.
[0168] Further, in the Laplacian component image data having low
resolution levels, density levels (pseudo contour) are more likely
to be noticed by emphasizing a very small slope. Due to this, the
manipulation content of FIG. 12 or 13, in which the manipulation
amounts are reduced, is applied to the pixels whose absolute values
of the Laplacian component image data are small.
[0169] As a result, according to the seventh embodiment, it becomes
possible to emphasize a signal component (contrast actually
existing in an object to be imaged) in the luminance (Y) component
image data without reducing the signal component more than
necessary while preventing the emphasis of a noise component and
the occurrence of the density levels (pseudo contour). That is, in
the image processing of manipulating the pixel values of the
multi-resolution image data, it becomes possible to intend to
improve image quality.
[0170] Although the invention has been described with respect to
specific embodiments for a complete and clear disclosure, the
appended claims are not to be thus limited but are to be construed
as embodying all modifications and alternative constructions that
may occur to one skilled in the art that fairly fall within the
basic teachings herein set forth.
[0171] The present application is based on and claims the benefit
of priority of Japanese Patent Application No. 2014-170387 filed
Aug. 25, 2014, the entire contents of which are hereby incorporated
herein by reference.
REFERENCE SIGNS LIST
[0172] 100 image processing apparatus [0173] 110 image processing
program [0174] 121 adjustment ratio DB [0175] 122 image
manipulation DB [0176] 320 luminance component processing section
[0177] 321 multi-resolution image generation section [0178] 322
multi-resolution image manipulation section [0179] 323
reconstruction section [0180] 330 adjustment section [0181] 340
color composition section [0182] 822 multi-resolution image
manipulation section [0183] 823 reconstruction section [0184] 830
adjustment section
CITATION LIST
Patent Literature
[0185] [PTL 1] Japanese Laid-open Patent Publication No.
2007-142670
[0186] [PTL 2] Japanese Laid-open Patent Publication No.
2014-068330
* * * * *